https://github.com/philipphager/ultr-cm-vs-ips

Source code for our paper "Contrasting Neural Click Models and Pointwise IPS Rankers".
https://github.com/philipphager/ultr-cm-vs-ips

clickmodels inverse-propensity-score ltr unbiased-learning-to-rank

Last synced: 2 months ago
JSON representation

Source code for our paper "Contrasting Neural Click Models and Pointwise IPS Rankers".

• Host: GitHub
• URL: https://github.com/philipphager/ultr-cm-vs-ips
• Owner: philipphager
• License: mit
• Created: 2022-07-26T12:33:20.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2023-04-17T13:49:28.000Z (over 1 year ago)
• Last Synced: 2023-04-26T11:01:30.498Z (over 1 year ago)
• Topics: clickmodels, inverse-propensity-score, ltr, unbiased-learning-to-rank
• Language: Jupyter Notebook
• Homepage:
• Size: 7.75 MB
• Stars: 7
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Contrasting Neural Click Models and Pointwise IPS Rankers

Source code for our paper `Contrasting Neural Click Models and Pointwise IPS Rankers`.

## Data

The project automatically downloads public datasets to `~/.ltr-datasets/` on first execution.

## Installation

1. Install dependencies using conda: `conda env create -f environment.yaml`

2. Activate environment: `conda activate ultr-cm-vs-ips`

3. Run experiments in the `/scripts directory`, e.g.: `./scripts/mslr_dataset_size.sh`

## Reference

```

@inproceedings{Hager2023Contrasting,

  author    = {Philipp Hager and Maarten de Rijke and Onno Zoeter},

  title     = {Contrasting Neural Click Models and Pointwise IPS Rankers},

  booktitle = {ECIR 2023: 45th European Conference on Information Retrieval},

  publisher = {Springer},

  year      = {2023},

}

```

## License

This project uses the [MIT license]().